From Customer Service to Code – everything is being described as a silver pill in the race to automate everything. The story is mock: AI tools that can write entire applications, smooth engineering teams and reduce the need for expensive human developers along with hundreds of other jobs.
But as a technician, from my point of view that spends every day within the data and workflows of real companies, the hype is not similar.
I have worked with industry leaders like General Electric, Walt Disney Company and Harvard Medical School to improve my data and AI infrastructure, and what I have learned is here. Changing humans with AI in most jobs is still an idea on the horizon.
I worry that we are thinking ahead. In the past two years, More than a quarter Programming jobs have disappeared. Mark Zuckerberg Announced He is planning to replace many of the Meta’s coders with AI.
But, surprisingly, Bill Gates and Sam Altman have both Publicly noted. Against the replacement of coders.
Right now, we should not trust AI tools to successfully replace jobs in a tech or business. The reason for this is that AI knows that what he has seen is naturally limited – and what he has seen in the tech world is the boilerplate.
Generative AI models are trained on major datases, which usually come in two main types: publicly available data (from open internet), or proprietary or licensed data (manufactured at home, or purchased from the third party).
Easy work, such as creating a basic website or forming a template app, are easy to win for generative models. But when it comes to writing a sophisticated, proprietary infrastructure code that gives power to companies like Google or Strip, there is a problem: this code is not available in public reserves. It is locked inside the walls of corporations, which is inaccessible to training data and is often written with decades of experience by engineers.
Right now, Ai Cannot cause the reason Now yourself. And there is no dignity. This is just a pattern imitation. A friend of mine in the tech world once described the big language model (llm) as a "Really good guess."
Think of AI today as a member of the junior team – help first or easy to make plans. But like any junior, it also needs to be monitored. In programming, for example, when I have achieved a 5X improvement for simple coding, I have found that more complex AI -made code is often needed and more than writing the code is often needed than writing code.
Finding flaws You You still need senior professionals with deep experience, and to understand the nuances that these flaws may be at risk of six months from now.
This does not mean that AI should not be placed in the workplace. But it is premature to host a man and AI tools instead of programmers or accountants or marketers. We still need senior level people in these jobs, and we need to train people in junior level jobs so they can one day technically be able to play a more complicated role.
AI’s target in tech and business should not be about removing humans from the loop. I’m not saying that because I am afraid that Ai will take my job. I am saying that because I have seen how dangerous it can be to trust AI at this stage.
Business leaders, regardless of what industry they are in, should be aware of: while AI promises cost savings and small teams, but the benefits of these performance can be beneficial. You may be relied on AI to perform more at the junior level of work, but not to complete more sophisticated projects.
Ai is fast. Humans are smart. There is a big difference. The sooner we change the conversation by strengthening them from the place of human beings, the more we will get the benefits of AI.
Derek Chang’s founder partner is Strats data.